Welcome to the newest edition of ID Tech’s AI Update. Here’s the latest big news on the shifting landscape of AI and identity technology:
OpenAI is considering demanding a cut of earnings from scientific breakthroughs enabled by its AI technology. CFO Sarah Friar openly discussed such a possibility on the firm’s podcast: “Let’s say in drug discovery, if we licensed our technology, you have a breakthrough,” she said. “The drug takes off, and we get a licensed portion of all its sales. It’s great alignment for us with our customer.” It’s the latest sign of OpenAI’s efforts to further monetize its capex-heavy tech, after last week’s news that the cheapest and free versions of ChatGPT would soon include third-party ads.
OpenAI rival Anthropic has cut its internal 2025 gross profit margin projection from about 50 percent to 40 percent. While Anthropic is projecting a roughly 12x increase in revenues year-over-year, the startup has seen unexpectedly high inference costs on Google and Amazon infrastructure – these are the costs of actually processing customer queries and prompts.
Intel is struggling to meet demand for server-side CPUs in AI data centers. In its latest quarterly update, the company actually reported a four percent drop in revenues year-over-year, but that’s mainly due to its struggling PC CPU business. Its data center and AI division meanwhile, saw growth of nine percent and its Intel Foundry saw growth of four percent. “In the short term, I’m disappointed that we are not able to fully meet the demand in our markets,” said CEO Lip-Bu Tan.
Tesla expects to start selling its Optimus humanoid robots to customers by 2027, according to CEO Elon Musk. Speaking at the World Economic Forum in Davos, Musk said that Optimus is already performing “simple tasks” in factories, and that it’s expected to be capable of more complex tasks later this year, paving the way for sales to consumers.
Apple is reportedly working on a small, screenless, wearable AI “pin” that would use cameras and mics to understand the user’s surroundings and interact with the user via voice. It’s expected to be a flat, circular device similar in size to an AirTag, and to leverage a more advanced Siri chatbot. But its future remains up in the air; it might come out next year, or it could be scrapped entirely. Apple has not officially commented on the project.
Generative video specialist Runway says most people can’t tell the difference between AI-generated video and camera-captured video. To promote its latest base model, Gen-4.5, it polled a random sampling of 1,043 participants, and showed them 20 videos – 10 AI-generated, 10 real – in randomized order, asking them to judge which were generated by AI. According to Runway, only 9.5 percent “achieved statistically significant accuracy (≥15/20 correct, p < 0.05, binomial test)”. Overall accuracy was only a little above chance, at 57.1 percent. “Performance was similar on real (58.0%) and generated (56.1%) videos, indicating no systematic detection strategy.”
The chatbot’s take: We wanted a little more context around Runway’s claims:
–
By Alex Perala
